Scrutable

Woodchopper wrote: Mon Apr 06, 2020 12:39 pm

shpalman wrote: Mon Apr 06, 2020 11:12 am This morning I tried setting up a SEIQR model in a spreadsheet. (Susceptible - Exposed - Infectious - Quarantined - Recovered, in which I'm assuming a certain probability of an infection getting detected and quarantined, alternatively it could recover without being detected).

I've realized a thing: if the number of new infections per day is going down, it does not mean that the reproduction number has dropped below 1 and that the infections are tending to zero. It just means you reduced the contagion parameter β. If the number of infectious people (I) is climbing anyway, then once the product βI reaches the value it had before, you'll get the same infection rate you had before. This seems to have happened a few times in Italy already: "yay it's going down no wait it's gone up again".

This will carry on until a substantial fraction of the susceptible population has been infected. Or you really do manage to get the reproduction number to below 1. But if it's not below 1 now in Italy I struggle to think of what more can be done.

It looks like mortality is declining now in Italy, Spain, Belgium, Austria. So it looks like the lockdown measures implemented there about 3-4 weeks ago are having an effect. So either the lockdowns have got R to below 1 on average at a national level, or at least the lockdowns have protected the most vulnerable members of society.

The number of confirmed cases is not the same as the number of actual cases. It depends on how many infected people are detected.

The epidemics in both countries may be growing at a similar rate now, but early on the UK had more diagnosed cases than Italy. Italian numbers shot up on 23 February, leading scientists to think there was a period when the virus was spreading without being detected.

That gave less room for measures like tracing contacts of those who had fallen ill and isolating cases to slow the spread.

Professor of international public health Jimmy Whitworth says that put the health system "behind the curve" in controlling the epidemic.

Researchers also warn that the Italy virus testing system has become overwhelmed and is not keeping up with new cases. This means Italy's figures could be falling further behind the total number of actual cases.

lpm wrote: Mon Apr 06, 2020 7:14 pm But the UK has had 3 days of very low deaths. Maybe dying in homes or a weekend effect? Still, currently on track for a better outcome than we feared.

TopBadger wrote: Mon Apr 06, 2020 12:44 pm

jimbob wrote: Sun Apr 05, 2020 4:33 pm Should this bad graph be here?

Argh - that Y axis! Why did you post that... why!?

Squeak wrote: Mon Apr 06, 2020 1:35 pm Cautiously, Australia just might be managing this ok, so far, touch wood, and all similar invocations of hope.

A couple of weeks ago, we were doubling cases every three-and-a-bit days. Now it's ten days. Our testing rates are on par with Germany and Italy but only 2.1% of tests are coming back positive so we're not just testing the obvious cases but we're probably picking up most of what's out there. That we've stopped the exponential growth in new cases and they or mortality rate seems to be around 0.2% add more support to the idea that there aren't large pockets of mild but infectious disease that our testing systems are missing.

Some of this is dumb luck of geography that gave us some time to slowly organise a social/political response and sprawling cities that make social distancing easier. But maybe, just maybe, if we keep doing what we're doing, we can keep this bubbling at manageable levels until we get a vaccine.

Have I added enough caveats yet? Obvs new data tomorrow night undermine all of this or Easter beach holidays might screw it up, even if the current trends is right.

sTeamTraen wrote: Tue Apr 07, 2020 12:01 am What would the effect on observance of the not-really-a-lockdown be if the Prime Minister's outcome was to demonstrate the vulnerability of the highest in the land to the virus? Would it change anything?

Woodchopper wrote: Tue Apr 07, 2020 5:22 am Some of the tories on my social media feeds have stopped stating that’s it’s all media hysteria and Covid is no worse than flu.

bmforre wrote: Tue Apr 07, 2020 5:49 am

Woodchopper wrote: Tue Apr 07, 2020 5:22 am Some of the tories on my social media feeds have stopped stating that’s it’s all media hysteria and Covid is no worse than flu.

Don't they any longer trust Trump, Bolsonaro and Even Higher powers?

lpm wrote: Mon Apr 06, 2020 7:14 pm But the UK has had 3 days of very low deaths. Maybe dying in homes or a weekend effect? Still, currently on track for a better outcome than we feared.

raven wrote: Mon Apr 06, 2020 11:48 pm

lpm wrote: Mon Apr 06, 2020 7:14 pm But the UK has had 3 days of very low deaths. Maybe dying in homes or a weekend effect? Still, currently on track for a better outcome than we feared.

Who knows. I'm a bit worried that this is true - https://www.theguardian.com/world/2020/ ... p-too-late

The initial advice to stay home and not phone 111 unless you got worse wasn't totally clear to some people, who thought they were to stay home full stop, no help coming type of deal. And there'll be people who just don't want to go near a hospital right now, so I can believe people are waiting too long.

EACLucifer wrote: Sun Apr 05, 2020 12:22 pm Cloth masks seem a very sensible way to get some degree of mass mask wearing without drawing on stocks of PPE needed by healthcare and other key workers. Yes, the government should have prepared, given the warning they had, but they didn't - as citizens, we can complain, or we can complain and also fix the f.cking problem.

I've started work on trying to make masks for everyone on my street. The test example I ran up first seems to work fine, and is comfortable enough. It's based on a design requested by a health provider in the States, although with the elastic patterned after a painter's mask instead, for greater comfort.

Unfortunately, I - unrelatedly - hurt my hip badly enough to leave me stuck in bed, but once I'm back to work, would people here be interested in an illustrated guide to making them?

Woodchopper wrote: Tue Apr 07, 2020 7:24 am Well lower, in any other context it would be bad. Its actually two days in a row of markedly lower mortality (5 and 6 April), and one day of lower confirmed cases. I think its too early to declare a trend, but if mortality continues to decline for another few more days then we can start to be hopeful.

bob sterman wrote:In China they took mildly symptomatic cases out of their homes, and put them into isolation facilities (not full hospitals) for monitoring and to prevent household transmission.

raven wrote: Tue Apr 07, 2020 12:35 pm

Woodchopper wrote: Tue Apr 07, 2020 7:24 am Well lower, in any other context it would be bad. Its actually two days in a row of markedly lower mortality (5 and 6 April), and one day of lower confirmed cases. I think its too early to declare a trend, but if mortality continues to decline for another few more days then we can start to be hopeful.

I'm dubious. If weekend deaths aren't getting added until mid-week, if we're missing a bunch of community deaths... Actually, why is it the deaths aren't back-dated to the date they occured? Wouldn't that help the people monitoring this get a more accurate picture of what's happening?

bob sterman wrote:In China they took mildly symptomatic cases out of their homes, and put them into isolation facilities (not full hospitals) for monitoring and to prevent household transmission.

Yes, that's the bit our strategy is missing. And, having got Johnson's letter this morning, complete with a diagram explaining how long a 4-person household should isolate as one after another get it.... I mean, it shows person no.3, who's symptomless, out of isolation while person no.4 is symptomatic. Surely that's not great?